Abstract

Code-Mixing (CM) is a natural phenomenon observed in many multilingual societies and is becoming the preferred medium of expression and communication in online and social media fora. In spite of this, current Question Answering (QA) systems do not support CM and are only designed to work with a single interaction language. This assumption makes it inconvenient for multi-lingual users to interact naturally with the QA system especially in scenarios where they do not know the right word in the target language. In this paper, we present WebShodh - an end-end web-based Factoid QA system for CM languages. We demonstrate our system with two CM language pairs: Hinglish (Matrix language: Hindi, Embedded language: English) and Tenglish (Matrix language: Telugu, Embedded language: English). Lack of language resources such as annotated corpora, POS taggers or parsers for CM languages poses a huge challenge for automated processing and analysis. In view of this resource scarcity, we only assume the existence of bi-lingual dictionaries from the matrix languages to English and use it for lexically translating the question into English. Later, we use this loosely translated question for our downstream analysis such as Answer Type(AType) prediction, answer retrieval and ranking. Evaluation of our system reveals that we achieve an MRR of 0.37 and 0.32 for Hinglish and Tenglish respectively. We hosted this system online and plan to leverage it for collecting more CM questions and answers data for further improvement.

Abstract

Word embeddings learned from text corpus can be improved by injecting knowledge from external resources, while at the same time also specializing them for similarity or relatedness. These knowledge resources (like WordNet, Paraphrase Database) may not exist for all languages. In this work we introduce a method to inject word embeddings of a language with knowledge resource of another language by leveraging bilingual embeddings. First we improve word embeddings of German, Italian, French and Spanish using resources of English and test them on variety of word similarity tasks. Then we demonstrate the utility of our method by creating improved embeddings for Urdu and Telugu languages using Hindi WordNet, beating the previously established baseline for Urdu.

Abstract

This paper presents DILTON, a system which solves simple arithmetic word problems. DILTON first predicts the operation that is to be performed (’-’,’+’,’*’,’/’) through a deep neural network based model and then uses it to generate the answer. DILTON divides the question into two parts - world state and query as shown in Figure 1. The world state and the query are processed separately in two different networks and finally the networks are merged to predict the final operation. DILTON learns to predict operations with 8.81 % in a corpus of primary school questions. With simple similarity between the contexts of quantities appearing in the problem and the question text, we are able to identify 92.25 % of relevant quantities and solve 81% of the questions. Our code and data is publicly available

Abstract

We present a language independent, unsupervised approach for transforming word embeddings from source language to target language using a transformation matrix. Our model handles the problem of data scarcity which is faced by many languages in the world and yields improved word embeddings for words in the target language by relying on transformed embeddings of words of the source language. We initially evaluate our approach via word similarity tasks on a similar language pair - Hindi as source and Urdu as the target language, while we also evaluate our method on French and German as target languages and English as source language. Our approach improves the current state of the art results - by 13% for French and 19% for German. For Urdu, we saw an increment of 16% over our initial baseline score. We further explore the prospects of our approach by applying it on multiple models of the same language and transferring words between the two models, thus solving the problem of missing words in a model. We evaluate this on word similarity and word analogy tasks.

Abstract

We present a language independent, unsupervised method for building word embeddings using morphological expansion of text. Our model handles the problem of data sparsity and yields improved word embeddings by relying on training word embeddings on artificially generated sentences. We evaluate our method using small sized training sets on eleven test sets for the word similarity task across seven languages. Further, for English, we evaluated the impacts of our approach using a large training set on three standard test sets. Our method improved results across all languages.

Abstract

We propose the use of the Newsela corpus for Complex Word Identification, a sub-problem of Lexical Simplification and conduct an empirical evaluation by comparing it with benchmark corpora previously employed for this task. Our experiments suggest that the proposed corpus is effective for Complex Word Identification, thus helping Lexical Simplification.

Abstract

When it comes to computational language generation systems, humour is a relatively unexplored domain, especially more so for Hindi (or rather, for most languages other than English). Most researchers agree that a joke consists of two main parts-the setup and the punchline, with humour being encoded in the incongruity between the two. In this paper, we look at Dur se Dekha jokes, a restricted domain of humorous three liner poetry in Hindi. We analyze their structure to understand how humour is encoded in them and formalize it. We then develop a system which is successfully able to generate a basic form of these jokes.

Abstract

Sexism is prevalent in today’s society, both offline and online, and poses a credible threat to social equality with respect to gender. According to ambivalent sexism theory (Glick and Fiske, 1996), it comes in two forms: Hostile and Benevolent. While hostile sexism is characterized by an explicitly negative attitude, benevolent sexism is more subtle. Previous works on computationally detecting sexism present online are restricted to identifying the hostile form. Our objective is to investigate the less pronounced form of sexism demonstrated online. We achieve this by creating and analyzing a dataset of tweets that exhibit benevolent sexism. By using Support Vector Machines (SVM), sequence-to-sequence models and FastText classifier, we classify tweets into ‘Hostile’,‘Benevolent’or ‘Others’ class depending on the kind of sexism they exhibit. We have been able to achieve an F1-score of 87.22% using FastText classifier. Our work helps analyze and understand the much prevalent ambivalent sexism in social media.

Abstract

This work presents a corpus of Bollywood song lyrics and its metadata, annotated with sentiment polarity. We call this BolLy. It contains lyrics of 1055 songs ranging from those composed in the year 1970 to the most recent ones. This dataset is of utmost value as all the annotation is done manually by three annotators and this makes it a very rich dataset for training purposes. In this work, we describe the creation and annotation process, content, and the possible uses of the dataset. As an experiment, we have built a basic classification system to identify the emotion polarity of the song based solely on the lyrics and this can be used as a baseline algorithm for the same. BolLy can also be used for studying code-mixing with respect to lyrics.

Abstract

Sentiment Analysis is one of the most active research areas in natural language processing and an extensively studied problem in data mining, web mining and text mining for English language. With the proliferation of social media these days, data is widely increasing in regional languages along with English. Telugu is one such regional language with abundant data available in social media, but it’s hard to find a labeled training set as human annotation is time-consuming and cost-ineffective. To address this issue, in this paper the practicality of active learning for Telugu sentiment analysis is investigated. We built a hybrid approach by combining different query selection strategy frameworks to increase more accurate training data instances with limited labeled data. Using a set of classifiers like SVM, XGBoost, and Gradient Boosted Trees (GBT), we achieved promising results with minimal error rate.